Bug#864802: unblock: squashfs-tools/1:4.3-4

[GCS]

Package: release.debian.org
User: release.debian@packages.debian.org
Usertags: unblock

Hi Release Team,

There are two data corruption bug in squashfs-tools that fixed in the
last upload for Sid. I let it age seven days even if the fixes quite
straightforward.

The first is due to a rare race condition of filesystem
finalization[1] and the fix is to hold the tread lock longer until the
write buffer is put in line on the queue.
+-  pthread_mutex_unlock(_mutex);
+   queue_put(to_writer, write_buffer);
++  pthread_mutex_unlock(_mutex);

The second is a 2 GB limit in file size in certain conditions as one
place used a wrong (limited in size) variable type. As such, the fix
is the following.
+-  int file_size = inode->buf.st_size;
++  off_t file_size = inode->buf.st_size;
An Endless OS developer also confirmed[2] that the fix is correct.

Full debdiff is attached.

Thanks for consideration,
Laszlo/GCS
[1] 
https://github.com/plougher/squashfs-tools/commit/de03266983ceb62e5365aac84fcd3b2fd4d16e6f
[2] https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=788185#37
diff -Nru squashfs-tools-4.3/debian/changelog squashfs-tools-4.3/debian/changelog
--- squashfs-tools-4.3/debian/changelog	2015-10-21 20:03:07.0 +
+++ squashfs-tools-4.3/debian/changelog	2017-06-07 17:47:58.0 +
@@ -1,3 +1,12 @@
+squashfs-tools (1:4.3-4) unstable; urgency=medium
+
+  * Backport patch to fix rare race in fragment waiting in filesystem
+finalisation.
+  * Backport fix for 2GB-limit of the is_fragment(...) function
+(closes: #788185).
+
+ -- Laszlo Boszormenyi (GCS)   Wed, 07 Jun 2017 17:47:58 +
+
 squashfs-tools (1:4.3-3) unstable; urgency=low
 
   * Use patch from upstream BTS to support LZMA magics (closes: #802446).
diff -Nru squashfs-tools-4.3/debian/patches/0006-uptream-fix-race.patch squashfs-tools-4.3/debian/patches/0006-uptream-fix-race.patch
--- squashfs-tools-4.3/debian/patches/0006-uptream-fix-race.patch	1970-01-01 00:00:00.0 +
+++ squashfs-tools-4.3/debian/patches/0006-uptream-fix-race.patch	2017-06-07 17:47:58.0 +
@@ -0,0 +1,54 @@
+commit de03266983ceb62e5365aac84fcd3b2fd4d16e6f
+Author: Phillip Lougher 
+Date:   Thu Sep 18 01:28:11 2014 +0100
+
+mksquashfs: fix rare race in fragment waiting in filesystem finalisation
+
+Fix a rare race condition in fragment waiting when finalising the
+filesystem.  This is a race condition that was initially fixed in 2009,
+but inadvertantly re-introduced in the latest release when the code
+was rewritten.
+
+Background:
+
+When finalising the filesystem, the main control thread needs to ensure
+all the in-flight fragments have been queued to the writer thread before
+asking the writer thread to finish, and then writing the metadata.
+
+It does this by waiting on the fragments_outstanding counter.  Once this
+counter reaches 0, it synchronises with the writer thread, waiting until
+the writer thread reports no outstanding data to be written.
+
+However, the main thread can race with the fragment deflator thread(s)
+because the fragment deflator thread(s) decrement the fragments_outstanding
+counter and release the mutex before queueing the compressed fragment
+to the writer thread, i.e. the offending code is:
+
+fragments_outstanding --;
+pthread_mutex_unlock(_mutex);
+queue_put(to_writer, write_buffer);
+
+In extremely rare circumstances, the main thread may see the
+fragments_outstanding counter is zero before the fragment
+deflator sends the fragment buffer to the writer thread, and synchronise
+with the writer thread, and finalise before the fragment has been written.
+
+The fix is to ensure the fragment is queued to the writer thread
+before releasing the mutex.
+
+Signed-off-by: Phillip Lougher 
+
+diff --git a/squashfs-tools/mksquashfs.c b/squashfs-tools/mksquashfs.c
+index 87b7d86..f1fcff1 100644
+--- a/squashfs-tools/mksquashfs.c
 b/squashfs-tools/mksquashfs.c
+@@ -2445,8 +2445,8 @@ void *frag_deflator(void *arg)
+ 			write_buffer->block = bytes;
+ 			bytes += compressed_size;
+ 			fragments_outstanding --;
+-			pthread_mutex_unlock(_mutex);
+ 			queue_put(to_writer, write_buffer);
++			pthread_mutex_unlock(_mutex);
+ 			TRACE("Writing fragment %lld, uncompressed size %d, "
+ "compressed size %d\n", file_buffer->block,
+ file_buffer->size, compressed_size);
diff -Nru squashfs-tools-4.3/debian/patches/0007-fix-2GB-limit-in-mksquashfs.patch squashfs-tools-4.3/debian/patches/0007-fix-2GB-limit-in-mksquashfs.patch
--- squashfs-tools-4.3/debian/patches/0007-fix-2GB-limit-in-mksquashfs.patch	1970-01-01 00:00:00.0 +
+++ 

request not to remove the package as people asked

[Thierry Vilmart]

Jonas Smedegaard wrote:

> Exactly: Packages poorly _maintained_ should be removed.  E.g. npm!
> My point in previous post was that focusing only on the workload for 
> _initial_ packaging masks the actual real workload, which is being 
> discussed here!
> - Jonas

I agree. If it is too complicated to maintain nodejs and npm packages, they 
should be removed for security reasons. There could be a vulnerability in the 
nodejs web server.

Serious professionals will have to install tar balls from the official web 
pages.

Thierry Vilmart

Re: d-i-netboot-images package outdated (was Re: Debian Installer Stretch RC 5 release)

[Andrew M.A. Cater]

On Wed, Jun 14, 2017 at 08:45:47AM +0200, Cyril Brulebois wrote:
> Holger Levsen  (2017-06-13):
> > On Tue, Jun 13, 2017 at 10:19:17AM +0200, Cyril Brulebois wrote:
> > > Known bugs in this release
> > > ==
> > [...] 
> > > See the errata[2] for details and a full list of known issues.
> > 
> > https://tracker.debian.org/pkg/debian-installer-netboot-images hasn't seen 
> > an
> > update in a while (and thus is unusuable due to kernel version skew), is it
> > on your collective radar to update the package til Saturday?
> > 
> > (Debian Edu uses that packages to support installation via PXE out of the 
> > box.)
> > 
> > Shall I file an RC bug to make the problem more visible and known?
> 
> As mentioned in the announce: We're doing another d-i upload for the
> release anyway.
> 
> Not that the latest d-i-n-i upload was a bad thing. It's just going to
> be superseded.
> 
> (Last two items on my r0 checklist:
> d-i
> d-i-n-i after d-i
> 

Seeing the post on PXE for UEFI on planet.debian.org and noting that
you're planning another d-i release.

Is there any chance of putting in the symlink in d-i that will link 
bootnetx64.efi in the same way as pxelinux as below

Also in netboot.tar.gz similarly

bootnetx64.efi -> debian-installer/amd64/bootnetx64.efi

This is exactly the way that pexlinux.0 and pxelinux.cfg are already
linked and would be a trivial change that would allow UEFI booting
more readily.

Thank you for your consideration

Andy C.


> … and Karsten has committed the d-i fixes for i2c, so we should be go to
> go for an upload; need to catch up with some more mails before that
> though.)
> 
> 
> KiBi.

Bug#864783: RM: aiccu/20070115-17

[Axel Beckert]

Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: rm
Control: severity 863720 serious
Control: found 863720 20070115-14

Hi,

the sunset of SixXS has passed, see https://bugs.debian.org/863720 for
details and discussion.

And unfortunately, neither

* SixXS changed their mind, nor
* did SixXS open-source the server implementation, nor
* did someone else come up with an AYIYA(*) server implementation or a
  comparable service.

There though seems to be one ISP though which uses aiccu to setup
tunnels for his own customers, but also only for them as the
authentication seems to be based on the MAC address from where the
TIC(**) request comes from. See
https://en.wikipedia.org/wiki/AICCU#Usage and
http://n6.netbox.cz/mediawiki/index.php/AICCU (the latter is in Czech
language only.) I also doubt that they've implemented an AYIYA server
for the tunnel and suspect they use other tunnel types.

https://en.wikipedia.org/wiki/List_of_IPv6_tunnel_brokers also lists
AARNet as existing tunnel broker and at least in the past this broker
was supported by aiccu officially. (See
https://web.archive.org/web/20061026034634/http://www.sixxs.net/tools/aiccu/brokers/.)
And http://broker.aarnet.net.au/ indeed still exists. But then again
they're listed as "broken" on Wikipedia.

And the TIC server implementation I found on CPAN
(https://metacpan.org/pod/Net::SixXS::TIC::Server) is rather esoteric
respectively academic. While it seems to be able to communicate with
aiccu, no according AYIYA server implementation showed up so you can
only use it with tunnel protocols where better clients exist anyways.

So the only production-mode way to still use aiccu is with Netbox.cz
while being one of their customers. But neither Mike (AFAIK) nor I are
customers of Netbox.cz.

Hence I must agree with a heavy heart that it indeed is better to
remove aiccu from Stretch as we can't really fully test it anymore for
e.g. stable security updates or so.

Other distributions have removed aiccu as well, e.g. OpenWRT at
https://github.com/openwrt/packages/commit/441f8a3e

So please remove aiccu from Stretch. There seem to be no reverse
dependencies and also no reverse build dependencies.

Mike and me agreed to keep it around in Debian Unstable for at least a
few more months in the hope that any of the above mentioned events
still happen. If neither of these events happen (within a year or so),
we'll likely let aiccu also be removed from Debian Unstable.

Accordingly I'm setting #863720 back to serious with this mail to keep
it out of testing for now.

I'll probably also open an RFA with the explicit mention that the only
known use case are customers of Netbox.cz.

*sigh*

(*)  AYIYA is aiccu's most used tunneling protocol and the only one
 suitable for NAT traversal and dynamic IPs. Unfortunately there's
 no publically available server implementation.

(**) TIC is the Tunnel Information and Control protocol, i.e. a
 protocol to gather the necessary information to setup tunnels.

Regards, Axel
-- 
 ,''`.  |  Axel Beckert , http://people.debian.org/~abe/
: :' :  |  Debian Developer, ftp.ch.debian.org Admin
`. `'   |  4096R: 2517 B724 C5F6 CA99 5329  6E61 2FF9 CD59 6126 16B5
  `-|  1024D: F067 EA27 26B9 C3FC 1486  202E C09E 1D89 9593 0EDE


signature.asc
Description: Digital signature

Bug#864770: jessie-pu: package libapache2-mod-perl2/2.0.9~1624218-2+deb8u2

[Niko Tyni]

Package: release.debian.org
Severity: normal
Tags: jessie
User: release.debian@packages.debian.org
Usertags: pu
X-Debbugs-Cc: libapache2-mod-pe...@packages.debian.org

The changes in apache2_2.4.10-10+deb8u8 related to CVE-2016-8743
caused libapache2-mod-perl2 to start failing its test suite, as
seen in #864316.

The attached debdiff fixes this by amending the test suite.
The changes are identical to those we made in stretch/sid for #849082.

Please let me know if it's OK to upload to jessie.

Thanks for your work,
-- 
Niko Tyni   nt...@debian.org
diff -Nru libapache2-mod-perl2-2.0.9~1624218/debian/changelog 
libapache2-mod-perl2-2.0.9~1624218/debian/changelog
--- libapache2-mod-perl2-2.0.9~1624218/debian/changelog 2015-11-15 
20:42:37.0 +0200
+++ libapache2-mod-perl2-2.0.9~1624218/debian/changelog 2017-06-14 
14:39:56.0 +0300
@@ -1,3 +1,10 @@
+libapache2-mod-perl2 (2.0.9~1624218-2+deb8u2) jessie; urgency=medium
+
+  * Patch the test suite for apache2_2.4.10-10+deb8u8 compatibility.
+(Closes: #864316)
+
+ -- Niko Tyni   Wed, 14 Jun 2017 14:39:56 +0300
+
 libapache2-mod-perl2 (2.0.9~1624218-2+deb8u1) jessie; urgency=medium
 
   * Apply upstream 2.0.9 patches fixing crashes in
diff -Nru 
libapache2-mod-perl2-2.0.9~1624218/debian/patches/440_http_syntax.patch 
libapache2-mod-perl2-2.0.9~1624218/debian/patches/440_http_syntax.patch
--- libapache2-mod-perl2-2.0.9~1624218/debian/patches/440_http_syntax.patch 
1970-01-01 02:00:00.0 +0200
+++ libapache2-mod-perl2-2.0.9~1624218/debian/patches/440_http_syntax.patch 
2017-06-14 14:34:26.0 +0300
@@ -0,0 +1,33 @@
+From 4a803fdb4c9eae8538293fe31c9222eecb6465be Mon Sep 17 00:00:00 2001
+From: Niko Tyni 
+Date: Fri, 23 Dec 2016 18:27:23 +0200
+Subject: [PATCH 1/2] Fix t/apache/read.t HTTP syntax for Apache 2.4.24
+ compatibility
+
+HTTP/1.1 RFC 7230, section 2.6. "Protocol Versioning" says the HTTP name
+is case sensitive. Starting with Apache 2.4.24, using lower case will
+make the server issue a 400 Bad request response, causing a test failure.
+
+https://tools.ietf.org/html/rfc7230#section-2.6
+
+Bug-Debian: https://bugs.debian.org/849082
+---
+ t/apache/read.t | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+diff --git a/t/apache/read.t b/t/apache/read.t
+index 83670c9..9f7f504 100644
+--- a/t/apache/read.t
 b/t/apache/read.t
+@@ -24,7 +24,7 @@ close $fh;
+ 
+ my $size = length $data;
+ 
+-for my $string ("POST $location http/1.0",
++for my $string ("POST $location HTTP/1.0",
+ "Content-length: $size",
+ "") {
+ my $line = "$string\r\n";
+-- 
+2.11.0
+
diff -Nru 
libapache2-mod-perl2-2.0.9~1624218/debian/patches/450_inject_header_line_terminators.patch
 
libapache2-mod-perl2-2.0.9~1624218/debian/patches/450_inject_header_line_terminators.patch
--- 
libapache2-mod-perl2-2.0.9~1624218/debian/patches/450_inject_header_line_terminators.patch
  1970-01-01 02:00:00.0 +0200
+++ 
libapache2-mod-perl2-2.0.9~1624218/debian/patches/450_inject_header_line_terminators.patch
  2017-06-14 14:34:34.0 +0300
@@ -0,0 +1,45 @@
+From d59229cf4f5b91ed58e25e27977e76f59096b72d Mon Sep 17 00:00:00 2001
+From: Niko Tyni 
+Date: Sat, 24 Dec 2016 23:07:28 +0200
+Subject: [PATCH 2/2] Fix in_bbs_inject_header line terminators for Apache
+ 2.4.24 compatibility
+
+rfc7230 3.5 says:
+
+  Although the line terminator for the start-line and header fields is
+   the sequence CRLF, a recipient MAY recognize a single LF as a line
+   terminator and ignore any preceding CR.
+
+Apache with strict enabled chooses not to implement the MAY.
+
+Author: Stefan Fritsch 
+Bug-Debian: https://bugs.debian.org/849082
+---
+ t/filter/TestFilter/in_bbs_inject_header.pm | 4 ++--
+ 1 file changed, 2 insertions(+), 2 deletions(-)
+
+diff --git a/t/filter/TestFilter/in_bbs_inject_header.pm 
b/t/filter/TestFilter/in_bbs_inject_header.pm
+index b09d6f9..5380c65 100644
+--- a/t/filter/TestFilter/in_bbs_inject_header.pm
 b/t/filter/TestFilter/in_bbs_inject_header.pm
+@@ -181,7 +181,7 @@ sub handler : FilterConnectionHandler {
+ 
+ if ($data and $data =~ /^POST/) {
+ # demonstrate how to add a header while processing other headers
+-my $header = "$header1_key: $header1_val\n";
++my $header = "$header1_key: $header1_val\r\n";
+ push @{ $ctx->{buckets} }, APR::Bucket->new($c->bucket_alloc, 
$header);
+ debug "queued header [$header]";
+ }
+@@ -199,7 +199,7 @@ sub handler : FilterConnectionHandler {
+ # we hit the headers and body separator, which is a good
+ # time to add extra headers:
+ for my $key (keys %headers) {
+-my $header = "$key: $headers{$key}\n";
++my $header = "$key: $headers{$key}\r\n";
+ push @{ $ctx->{buckets} }, APR::Bucket->new($c->bucket_alloc, 
$header);
+

Bug#864655: stretch-pu: package trafficserver/7.0.0-5+deb9u1

[Adrian Bunk]

On Wed, Jun 14, 2017 at 09:12:21AM +0200, Jean Baptiste Favre wrote:
> Hello Adrian,
> 
> Thanks for pointing this out.
> 
> I've little knowledge in ARM dev.
> I've searched for memory barrier implementations examples in debian
> packages source code, but I was clearly presumptuous claiming to fix the
> build properly.
> 
> Do you have any pointer about proper memory barriers implementations ?
>...

std::atomic_thread_fence is the proper C++11 way to do memory barriers.

> Regards,
> Jean Baptiste

cu
Adrian

-- 

   "Is there not promise of rain?" Ling Tan asked suddenly out
of the darkness. There had been need of rain for many days.
   "Only a promise," Lao Er said.
   Pearl S. Buck - Dragon Seed

Last last installation guide upload?

[Samuel Thibault]

Hello,

Cyril Brulebois, on mar. 13 juin 2017 09:09:31 +0200, wrote:
> Holger Wansing  (2017-06-12):
> > > Source: installation-guide
> > Maybe we want another upload for this?
> > We got a nearly fully up-to-date manual for Chinese (simplified) these
> > days. zh_CN builds fine now (after some fixups by victory).
> 
> I'm fine with such an upload to unstable; but I'll likely defer to the
> release team whether to accept it for r0. Since that's mainly about
> documentation, I suspect that should be fine, but release date is very
> close.
> 
> On a practical level, not sure what bad consequences we could have.
> Steve already had to remove doc from the xfce CD so that the rest fits.

I have uploaded it, so it's now up to d-release to decide :)

Since it's not a new translation, just an update, package size have not
increased, they actually have *decreased*, I guess that's because it's
now only chinese in the files.

Samuel

Bug#864757: stretch-pu: package galternatives/0.13.5+nmu4+deb9u1

[Boyuan Yang]

在 2017年6月14日星期三 +08 上午9:32:28，Adam D. Barratt 写道：
> On 2017-06-14 8:41, Boyuan Yang wrote:
> > 在 2017年6月14日星期三 +08 上午8:35:36，Adam D. Barratt 写道：
> > 
> >> Control: tags -1 + moreinfo
> >> 
> >> On 2017-06-14 8:04, Boyuan Yang wrote:
> >> > I just adopted this package and now trying to fix some long-standing
> >> > problems.
> >> > 
> >> > This updated version has a one-liner fix for the old and annoying bug
> >> > #325172,
> >> > which made the package half-unusable. It would be great if the bug
> >> > could be
> >> > fixed in a point release of Stretch.
> >> 
> >> According to the BTS, this bug still affects the version of
> >> galternatives in unstable - is that correct? If so then it would need
> >> resolving there first.
> > 
> > A fixed version for unstable has already been sponsored into DELAYED to
> > coordinate with Stretch release [1],
> 
> That wasn't mentioned in either this bug or #325172. We're not psychic.
> 
> > so take it easy.
> 
> I'm going to assume that got lost in translation a little, because at
> least to an en_GB native speaker that comment came across quite
> impolitely.

Sorry if there is any offence, I am not a native English speaker.

I will remove the "moreinfo" tag after the deferred upload get accepted into 
unstable.

Boyuan Yang

signature.asc
Description: This is a digitally signed message part.

Re: d-i-netboot-images package outdated (was Re: Debian Installer Stretch RC 5 release)

[Holger Levsen]

On Wed, Jun 14, 2017 at 08:45:47AM +0200, Cyril Brulebois wrote:
> > Shall I file an RC bug to make the problem more visible and known?
> As mentioned in the announce: We're doing another d-i upload for the
> release anyway.
 
til then, having up2date debian-installer-netboot-images packages allows
people to actually test them…

> Not that the latest d-i-n-i upload was a bad thing. It's just going to
> be superseded.

sure!


-- 
cheers,
Holger


signature.asc
Description: Digital signature

Bug#864757: stretch-pu: package galternatives/0.13.5+nmu4+deb9u1

[Adam D. Barratt]

On 2017-06-14 8:41, Boyuan Yang wrote:

在 2017年6月14日星期三 +08 上午8:35:36，Adam D. Barratt 写道：

Control: tags -1 + moreinfo

On 2017-06-14 8:04, Boyuan Yang wrote:
> I just adopted this package and now trying to fix some long-standing
> problems.
>
> This updated version has a one-liner fix for the old and annoying bug
> #325172,
> which made the package half-unusable. It would be great if the bug
> could be
> fixed in a point release of Stretch.

According to the BTS, this bug still affects the version of
galternatives in unstable - is that correct? If so then it would need
resolving there first.


A fixed version for unstable has already been sponsored into DELAYED to
coordinate with Stretch release [1],


That wasn't mentioned in either this bug or #325172. We're not psychic.


so take it easy.


I'm going to assume that got lost in translation a little, because at 
least to an en_GB native speaker that comment came across quite 
impolitely.


Regards,

Adam

Bug#864757: stretch-pu: package galternatives/0.13.5+nmu4+deb9u1

[Boyuan Yang]

在 2017年6月14日星期三 +08 上午8:35:36，Adam D. Barratt 写道：
> Control: tags -1 + moreinfo
> 
> On 2017-06-14 8:04, Boyuan Yang wrote:
> > I just adopted this package and now trying to fix some long-standing
> > problems.
> > 
> > This updated version has a one-liner fix for the old and annoying bug
> > #325172,
> > which made the package half-unusable. It would be great if the bug
> > could be
> > fixed in a point release of Stretch.
> 
> According to the BTS, this bug still affects the version of
> galternatives in unstable - is that correct? If so then it would need
> resolving there first.

A fixed version for unstable has already been sponsored into DELAYED to 
coordinate with Stretch release [1], so take it easy.

[1] https://bugs.debian.org/864702

--
Boyuan Yang

signature.asc
Description: This is a digitally signed message part.

Re: Accepted thermald 1.5.4-2.1 (source amd64) into testing-proposed-updates

[Adam D. Barratt]

Hi,

On 2017-06-13 16:03, Colin King wrote:

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA256

Format: 1.8
Date: Tue, 13 Jun 2017 15:23:11 +0100
Source: thermald
Binary: thermald
Architecture: source amd64
Version: 1.5.4-2.1
Distribution: stretch
Urgency: medium
Maintainer: Colin King 
Changed-By: Colin King 
Description:
 thermald   - Thermal monitoring and controlling daemon
Closes: 864707
Changes:
 thermald (1.5.4-2.1) stretch; urgency=medium


What was the intent of this upload? If it was to try and get the updated 
package into the release, then a) it's (far) too late and b) it should 
have been discussed, not just uploaded to t-p-u.


If, otoh, this was intended for p-u after the release then a) it's far 
too early and b) as above, please follow the usual p-u bug and 
discussion process, don't simply upload.


Regards,

Adam

Processed: Re: Bug#864757: stretch-pu: package galternatives/0.13.5+nmu4+deb9u1

[Debian Bug Tracking System]

Processing control commands:

> tags -1 + moreinfo
Bug #864757 [release.debian.org] stretch-pu: package 
galternatives/0.13.5+nmu4+deb9u1
Added tag(s) moreinfo.

-- 
864757: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=864757
Debian Bug Tracking System
Contact ow...@bugs.debian.org with problems

Bug#864757: stretch-pu: package galternatives/0.13.5+nmu4+deb9u1

[Adam D. Barratt]

Control: tags -1 + moreinfo

On 2017-06-14 8:04, Boyuan Yang wrote:
I just adopted this package and now trying to fix some long-standing 
problems.


This updated version has a one-liner fix for the old and annoying bug 
#325172,
which made the package half-unusable. It would be great if the bug 
could be

fixed in a point release of Stretch.


According to the BTS, this bug still affects the version of 
galternatives in unstable - is that correct? If so then it would need 
resolving there first.


Regards,

Adam

Re: Bug#864457: Bug#864536: missing kernel modules in D-I sd-card images

[Cyril Brulebois]

Karsten Merker  (2017-06-11):
> A patch against the kernel package is attached.  After that
> change is applied to the kernel package, we would then need to
> add the i2c-modules udeb to the installer builds for armhf.

So I've seen you pushed the changes in master for debian-installer. I'll
trigger another “daily” build for armhf, and the results should show up
at [1] in a few hours (we have a few crontab entries on dillon, so it's
not refreshed immediately).

 1. https://d-i.debian.org/daily-images/daily-build-overview.html


KiBi.


signature.asc
Description: Digital signature

Bug#864655: stretch-pu: package trafficserver/7.0.0-5+deb9u1

[Jean Baptiste Favre]

Hello Adrian,

Thanks for pointing this out.

I've little knowledge in ARM dev.
I've searched for memory barrier implementations examples in debian
packages source code, but I was clearly presumptuous claiming to fix the
build properly.

Do you have any pointer about proper memory barriers implementations ?
I'd really like to fix armel build and learn about this architecture.

Regards,
Jean Baptiste

On 13/06/2017 12:49, Adrian Bunk wrote:
> On Mon, Jun 12, 2017 at 09:58:07PM +0800, Aron Xu wrote:
>> ...
>> @@ -1,3 +1,11 @@
>> +trafficserver (7.0.0-5+deb9u1) UNRELEASED; urgency=medium
>> ...
>> +  * Add a patch to fix arm build
>> +
>> + -- Jean Baptiste Favre   Mon, 29 May 2017 14:45:52 
>> +0200
>> ...
>> ++  #if defined(__ARM_ARCH_4__) \
>> ++ || defined(__ARM_ARCH_4T__) \
>> ++ || defined(__ARM_ARCH_5__) \
>> ++ || defined(__ARM_ARCH_5E__)  \
>> ++ || defined(__ARM_ARCH_5T__) \
>> ++ || defined(__ARM_ARCH_5TE__) \
>> ++ || defined(__ARM_ARCH_5TEJ__) \
>> ++ || defined(__ARM_ARCH_6__) \
>> ++ || defined __ARM_ARCH_6J__  \
>> ++ || defined(__ARM_ARCH_6K__) \
>> ++ || defined(__ARM_ARCH_6Z) \
>> ++ || defined(__ARM_ARCH_6ZK__) \
>> ++ || defined(__ARM_ARCH_6T2__)
>> ++#if defined(__thumb__)
>> ++  // This is just a placeholder and almost certainly not sufficient.
>> ++  #define mb() __asm__ __volatile__ ("" : : : "memory");
>> ++  #define rmb() __asm__ __volatile__("" : : : "memory")
>> ++  #define wmb() __asm__ __volatile__("" : : : "memory")
>> ...
> 
> This would change the code from not building to being known broken on armel.
> 
> That's IMHO worse than before.
> 
> cu
> Adrian
> 




signature.asc
Description: OpenPGP digital signature

Bug#864757: stretch-pu: package galternatives/0.13.5+nmu4+deb9u1

[Boyuan Yang]

Package: release.debian.org
Severity: normal
Tags: stretch
User: release.debian@packages.debian.org
Usertags: pu

I just adopted this package and now trying to fix some long-standing problems.

This updated version has a one-liner fix for the old and annoying bug #325172, 
which made the package half-unusable. It would be great if the bug could be 
fixed in a point release of Stretch.

Thanks,
Boyuan Yang

diff -Nru galternatives-0.13.5+nmu4/debian/changelog 
galternatives-0.13.5+nmu4+deb9u1/debian/changelog
--- galternatives-0.13.5+nmu4/debian/changelog  2015-12-12 21:57:27.0 
+0800
+++ galternatives-0.13.5+nmu4+deb9u1/debian/changelog   2017-06-14 
10:17:30.0 +0800
@@ -1,3 +1,10 @@
+galternatives (0.13.5+nmu4+deb9u1) stretch; urgency=medium
+
+  * Adopt and switch maintainer information.
+  * Fix the bug which causes properties window blank. Closes: #325172
+
+ -- Boyuan Yang <073p...@gmail.com>  Wed, 14 Jun 2017 10:17:30 +0800
+
 galternatives (0.13.5+nmu4) unstable; urgency=medium
 
   * Non-maintainer upload.
diff -Nru galternatives-0.13.5+nmu4/debian/control 
galternatives-0.13.5+nmu4+deb9u1/debian/control
--- galternatives-0.13.5+nmu4/debian/control2015-12-12 21:57:27.0 
+0800
+++ galternatives-0.13.5+nmu4+deb9u1/debian/control 2017-06-14 
10:17:30.0 +0800
@@ -1,7 +1,9 @@
 Source: galternatives
 Section: admin
 Priority: optional
-Maintainer: Ryan Niebur 
+Maintainer: Debian Chinese Team 
+Uploaders:
+ Boyuan Yang <073p...@gmail.com>
 Build-Depends: debhelper (>= 7.2)
 Build-Depends-Indep: python, dh-python, intltool, gettext, python-glade2
 Standards-Version: 3.7.2
diff -Nru galternatives-0.13.5+nmu4/galternatives/main.py 
galternatives-0.13.5+nmu4+deb9u1/galternatives/main.py
--- galternatives-0.13.5+nmu4/galternatives/main.py 2015-12-12 
21:57:27.0 +0800
+++ galternatives-0.13.5+nmu4+deb9u1/galternatives/main.py  2017-06-14 
10:17:30.0 +0800
@@ -420,6 +420,7 @@
 
 def hide_details_cb (self, *args):
 self.details_window.hide ()
+return True
 
 def show_about_window_cb (self, *args):
 self.about_window.show_all ()


signature.asc
Description: This is a digitally signed message part.

Re: d-i-netboot-images package outdated (was Re: Debian Installer Stretch RC 5 release)

[Cyril Brulebois]

Holger Levsen  (2017-06-13):
> On Tue, Jun 13, 2017 at 10:19:17AM +0200, Cyril Brulebois wrote:
> > Known bugs in this release
> > ==
> [...] 
> > See the errata[2] for details and a full list of known issues.
> 
> https://tracker.debian.org/pkg/debian-installer-netboot-images hasn't seen an
> update in a while (and thus is unusuable due to kernel version skew), is it
> on your collective radar to update the package til Saturday?
> 
> (Debian Edu uses that packages to support installation via PXE out of the 
> box.)
> 
> Shall I file an RC bug to make the problem more visible and known?

As mentioned in the announce: We're doing another d-i upload for the
release anyway.

Not that the latest d-i-n-i upload was a bad thing. It's just going to
be superseded.

(Last two items on my r0 checklist:
d-i
d-i-n-i after d-i

… and Karsten has committed the d-i fixes for i2c, so we should be go to
go for an upload; need to catch up with some more mails before that
though.)


KiBi.


signature.asc
Description: Digital signature